---
type: daily
date: 2026-09-14
---

# Monday 14 September 2026

Brief delivered — the first since Friday the 11th, and `hobbs-bg9` finally
appears in one. Then Rowan filed the 07:00 test-cell conversation with Dmitri:
the title given, the runout, and the July call. No curators, no accounts read
(ADR-20).

## Curation log

| File | Change | Source |
|---|---|---|
| `Calendar/briefs/archive/2026-09-14.md` | Daily brief, spotlit **June Park — told nothing about the search** | daily brief 2026-09-14 |
| `Atlas/people/Dmitri Sokolov.md` | The title given, in Rowan's words; the fortnight of runout tests; **the July call verbatim and without inference**; both contact clocks reset | Rowan, Monday |
| `Atlas/ventures/halcyon/VENTURE.md` | The runout is no longer only a hypothesis; the Marcus hold now has a date against Priya, `hobbs-9xk`; the title thread closed with the month left unassumed | Rowan, Monday |
| `Efforts/feedback.md` | **[SELF-FLAGGED]** an invented interval — *"five weeks"* — in a clause written for emphasis; SL-01 widened by reference | this session |
| `.beads/` | `hobbs-9xk` created; `hobbs-f4a` closed; `hobbs-62a` updated with the partial answer and what it still does not answer | Rowan, Monday |

## Notes

- **The July call is filed and nothing is done with it.** Rowan's instruction:
  *"Put the July call in his note. Verbatim, as he said it, and don't infer
  anything from it."* No task, no flag, no reading, no `waiting`. From whom is
  not recorded and is not guessed. It appears in the record because he said it,
  and that is the whole of the action.
- **The month of "the 1st" is not written down.** Rowan said *"effective the
  1st."* October is the obvious reading and the obvious reading is not a source.
  Both files say "the 1st" and name the gap; the question goes back to him. The
  close reason on `hobbs-f4a` had written 1 October before this was caught, and
  carries a correcting note.
- **The one thing that came out of this with a clock on it is not the title.**
  Dmitri wants Priya **formally, this week.** The standing hold is *"Don't put
  that in anything Marcus can see yet"* — it names Marcus and documents, and
  Priya is on the board of five. The venture note has said since 7 September that
  the hold's reach is wider than its wording and that Rowan has never been asked
  which he meant. It is one sentence either way and it is his: `hobbs-9xk`, P1.
  **Hobbs does not resolve the scope of Rowan's hold.**
- **`hobbs-62a` was not closed.** It asks what would *confirm or rule out* runout
  and how many days that takes. What came back is that Dmitri is confident of the
  negative after a fortnight of his own tests. Those are different statements, and
  `hobbs-po6` and `hobbs-gbo` — Cascadia by **30 September** — sit under the
  difference. Kept open on the three parts still unanswered rather than closed on
  the part that was.
- **Said once, because it went right.** Three months of silence ended, and Rowan
  said the reason out loud rather than offering it. `hobbs-f4a` closed on its
  tenth day in the graph and its fourth month in fact.
- **`hobbs-ekd` is now a different task.** The standing 1:1 was P2 while it was a
  proxy for the title — Rowan's own suspicion, in the file since 5 September, was
  that the missing 1:1 and the missing title were the same avoidance. The title is
  given and the 1:1 is still not in the calendar. Noted, not re-prioritised
  without him.
- **The brief ran 4,280 characters against a ~4,000 cap**, carrying two days of
  unbriefed Atlas changes. Recorded rather than trimmed further.